A Mysterious Update: What Is Jakub Pachocki Promising?

Although OpenAI has not yet made any official announcement, internal information suggests this is more than just a minor bug fix. The company's chief scientist, Jakub Pachocki, reportedly told employees that GPT-5.6 will represent a "significant improvement" over the current GPT-5.5 standard.

What does that mean in practice? In the context of large language models (LLMs), "significant improvement" typically doesn't just mean better poetry-writing skills, but primarily involves the following technical parameters:

• Increased efficiency and speed: Reduced latency (delay) when generating responses.
• Better cost-effectiveness: Lower computational demands, which could lead to cheaper operations for businesses via the API.
• Safety guardrails: New mechanisms to limit hallucinations (fabricated facts) and better adherence to ethical guidelines.

For the average user, this means ChatGPT could be not only smarter at solving complex tasks, but also faster during real-time interaction.

The Battle for Dominance: OpenAI vs. Anthropic

OpenAI is under enormous pressure. The competition is not waiting around. Just a few days ago, Anthropic introduced its Claude Fable 5 model, which specializes in programming and complex office work.

In one of the most demanding AI tests, known as SWE-Bench Pro (a test focused on an AI's ability to solve real-world software problems), Claude Fable 5 achieved a success rate of over 80%. That is an extreme result in the field, placing Anthropic's model squarely at the top of developer rankings. If OpenAI fails to deliver GPT-5.6 with matching capabilities in "agentic" behavior (the AI's ability to autonomously complete chains of tasks), it risks losing market share among professional users.

For comparison, the current models rank in benchmark tests as follows:

Why Now? Business Strategy and IPO

The timing of the GPT-5.6 release is no coincidence. OpenAI is currently in a critical phase of transformation from a research organization into a global technology giant. According to available reports, the company is preparing materials for a potential initial public offering (IPO) within the next year.

Releasing a new, more powerful model shortly after the previous version serves to maintain strong growth momentum and demonstrate technological superiority to investors. At the same time, OpenAI needs to optimize its infrastructure – the more efficient the model (as GPT-5.6 is expected to be), the lower the operating costs during massive scaling.

Practical Impact for Czech Users and Businesses

What should someone in Prague, Brno, or Bratislava take away from all this?

1. Availability and the Czech Language

OpenAI models have long been very well localized. Both GPT-5.5 and the expected GPT-5.6 handle the Czech language at a top-tier level, including grammar and contextual nuances. For Czech companies that use AI for customer support or document automation, this means the new version could deliver even more accurate results in our language.

2. Price and Availability

For regular users, the subscription model will not change, but its efficiency likely will. Current prices for ChatGPT Plus are around $20 USD (approx. 470 CZK) per month. For developers and businesses using the API, the per-token price will be key. If GPT-5.6 is more efficient, there could be a reduction in the operating costs of AI applications in the Czech environment.

3. EU Regulation (AI Act)

As a European market, we are under the strict regulation of the EU AI Act. The speculated safety improvements in GPT-5.6 are crucial for us. The better the model meets safety and transparency standards, the easier its implementation will be for Czech companies that must comply with strict GDPR rules and European AI standards.

Conclusion: Is It Worth Waiting?

If you are a regular ChatGPT user, GPT-5.5 will likely suffice even now. However, if you are a programmer, data analyst, or business that builds its processes on AI, tracking the June release of GPT-5.6 is essential. The battle between OpenAI and Anthropic has just moved into a phase where it's no longer just about who knows the most, but who can translate that knowledge into action the fastest and most efficiently.